Method and apparatus for liveness detection

Abstract

There are provided a liveness detection method and device. The liveness detection method comprises: generating a random action instruction sequence including at least one random action instruction; sequentially sending a random action instruction in the random action instruction sequence; and determining whether the sequentially sent random action instruction in the random action instruction sequence is sequentially executed by a living body based on detection information of at least two sensors, wherein the at least two sensors comprise an image sensor and at least one non-image sensor; and determining that the liveness detection is succeeded if the sequentially sent random action instruction in the random action instruction sequence is sequentially executed by the living body. Accuracy of liveness detection can be improved by adopting the random action sequence and by combining images captured by the image sensor and information detected by the non-image sensor.

What is claimed is: 
     
       1. A liveness detection method performed by a processor, comprising:
 generating a random action instruction sequence including a plurality of random action instructions; 
 sequentially sending a random action instruction in the random action instruction sequence; 
 determining whether the sequentially sent random action instruction in the random action instruction sequence is sequentially executed by a living body based on detection information of at least two sensors, wherein the at least two sensors comprise an image sensor and at least one non-image sensor; 
 determining that the liveness detection is succeeded if the sequentially sent random action instruction in the random action instruction sequence is sequentially executed by the living body; and 
 determining that the liveness detection is failed if a currently sent random action instruction is determined to not be executed by the living body, wherein the random action instruction sequence at least comprises a third category of action instructions, for each third category of action instructions, the method further comprises: determining, by said processor, a first decision parameter based on a first detection information obtained from the image captured by the image sensor, determining, by said processor, a second decision parameter based on a second detection information generated by the at least one non-image sensor, and determining, by said processor, that the liveness detection is failed if the first decision parameter does not match with the second decision parameter, and wherein in the case that the currently sent random action instruction is not a last random action instruction in the random action instruction sequence, a next random action instruction is sent when the currently sent random action instruction is determined to be executed by the living body. 
 
     
     
       2. The liveness detection method according to  claim 1 , wherein the random action instruction sequence further comprises at least two categories of action instructions selected from a group comprising a first category of action instructions, a second category of action instructions, and a third category of action instructions,
 wherein the method further comprises:
 for each first category of action instructions, determining whether a currently sent action instruction of the first category is executed by the living body according to an image captured by the image sensor; 
 for each second category of action instructions, determining whether a currently sent action instruction of the second category is executed by the living body according to the information detected by the at least one non-image sensor; and 
 for each third category of action instructions, determining whether a currently sent action instruction of the third category is executed by the living body according to the image captured by the image sensor and the information detected by the at least one non-image sensor. 
 
 
     
     
       3. The liveness detection method according to  claim 2 , further comprising: processing information detected by the at least one non-image sensor to generate second detection information, and
 the liveness detection method further comprises at least one step of:
 recognizing an object in the image captured by the image sensor; 
 detecting image luminance in the image captured by the image sensor; 
 locating a face region in the image captured by the image sensor; 
 locating facial key points in the located face region; 
 extracting image texture information in the located face region, wherein the image texture information comprises at least one of skin texture and hair characteristic; and 
 obtaining a facial gesture based on the located facial key points, 
 wherein the first detection information obtained from the image captured by the image sensor comprises at least one of the facial gesture, the facial key points, the image texture information, the image luminance, and the object. 
 
 
     
     
       4. The liveness detection method according to  claim 3 , wherein if the first detection information comprises the facial gesture, the method further comprises:
 for each first category of action instructions, determining that the liveness detection is failed when the facial gesture does not match with the currently sent action instruction of the first category. 
 
     
     
       5. The liveness detection method according to  claim 1 , wherein the at least one non-image sensor comprises at least one of a light sensor, a distance sensor, an acceleration sensor and a gyroscope, and
 wherein the second detection information comprises at least one of light intensity, distance information, acceleration information, and gyroscope information. 
 
     
     
       6. The liveness detection method according to  claim 5 , wherein
 the first decision parameter comprises at least one of size and/or size variation of the object in the captured image, distance and/or distance variation among objects in the captured image, image luminance variation in the captured image, image luminance variation in the located facial region, the facial gesture, distance and/or distance variation among facial key points in the captured image, and the image texture information in the captured image, and 
 wherein the second decision parameter comprises at least one of light intensity and/or light intensity variation, a distance and/or distance variation of the object in the captured image relative to the distance sensor, a spatial position and/or spatial position variation of the object in the captured image relative to the acceleration sensor and/or the gyroscope. 
 
     
     
       7. A liveness detection device, comprising:
 an image sensor configured to capture an image; 
 at least one non-image sensor configured to detect information; 
 at least one storage means configured to store program instructions; and 
 at least one processor configured to execute the program instructions stored in the at least one storage means to:
 generate a random action instruction sequence including a plurality of random action instructions; 
 sequentially send a random action instruction in the random action instruction sequence; 
 process the image captured by the image sensor to generate first detection information; 
 process the information detected by at the least one non-image sensor to generate second detection information; 
 determine whether the sequentially sent random action instruction in the random action instruction sequence is sequentially executed by a living body based on the first detection information and the second detection information; 
 determine that the liveness detection is succeeded if the sequentially sent random action instruction in the random action instruction sequence is sequentially executed by the living body; and 
 determine that the liveness detection is failed if a currently sent random action instruction is determined to not be executed by the living body, 
 
 wherein the random action instruction sequence at least comprises a third category of action instructions, for each third category of action instructions, said at least one processor determines a first decision parameter based on the first detection information, determines a second decision parameter based on the second detection information, and determines that the liveness detection is failed if the first decision parameter does not match with the second decision parameter, 
 wherein in the case that the currently sent random action instruction is not a last action instruction in the random action instruction sequence, said at least one processor sends a next random action instruction when the currently sent random action instruction is determined to be executed by the living body. 
 
     
     
       8. The liveness detection device according to  claim 7 , wherein the random action instruction sequence comprises at least two categories of action instructions selected from a group comprising a first category of action instructions, a second category of action instructions and a third category of action instructions;
 wherein for each first category of action instructions, said at least one processor determines whether a currently sent action instruction of the first category is executed by the living body according to the first detection information; 
 for each second category of action instructions, said at least one processor determines whether a currently sent action instruction of the second category is executed by the living body according to the second detection information; and 
 for each third category of action instructions, said at least one processor determines whether the currently sent action instruction of the third category is executed by the living body according to the first detection information and the second detection information. 
 
     
     
       9. The liveness detection device according to  claim 8 , wherein the first detection information comprises at least one of facial gesture, facial key points, image texture information, image luminance, and an object recognized in the image captured by the image sensor. 
     
     
       10. The liveness detection device according to  claim 9 , wherein
 when the first detection information comprises the facial gesture, for each first category of action instructions, said at least one processor determines that the liveness detection is failed if the facial gesture does not match with the currently sent first category of action instruction. 
 
     
     
       11. The liveness detection device according to  claim 7 , wherein the at least one non-image sensor comprises at least one of a light sensor, a distance sensor, an acceleration sensor and a gyroscope, and
 wherein the second detection information comprises at least one of light intensity, distance information, acceleration information, and gyroscope information. 
 
     
     
       12. The liveness detection device according to  claim 11 , wherein
 the first decision parameter comprises at least one of size and/or size variation of an object in the captured image, distance and/or distance variation among objects in the captured image, image luminance variation in the captured image, image luminance variation in a located face region, the facial gesture, distance and/or distance variation among facial key points in the captured image, and image texture information in the captured image; and 
 wherein the second decision parameter comprises at least one of light intensity and/or light intensity variation, a distance and/or distance variation of the object in the captured image relative to the distance sensor, a spatial position and/or spatial position variation of the object in the captured image relative to the acceleration sensor and/or the gyroscope. 
 
     
     
       13. A non-transitory computer readable storage medium with program instructions recorded thereon, when being executed by a computer or a processor, the program instructions make the computer or the processor to:
 generate a random action instruction sequence including a plurality of random action instructions; 
 sequentially send a random action instruction in the random action instruction sequence; 
 process an image captured by an image sensor to generate a first detection information; 
 process information detected by at least one non-image sensor to generate a second detection information; 
 determine whether the sequentially sent random action instruction in the random action instruction sequence is sequentially executed by a living body based on detection information of at least two sensors; 
 determine that the living body is detected if the sequentially sent random action instruction in the random action instruction sequence is sequentially executed by the living body; and 
 determine that the liveness detection is failed if a currently sent random action instruction is determined to not be executed by the living body, 
 wherein the random action instruction sequence at least comprises a third category of action instructions, for each third category of action instructions, when being executed by a computer or a processor, the program instructions make the computer or the processor to determine a first decision parameter based on the first detection information, determines a second decision parameter based on the second detection information, and determine that the liveness detection is failed if the first decision parameter does not match with the second decision parameter, 
 wherein in the case that the currently sent random action instruction is not the last random action instruction in the random action instruction sequence, a next random action instruction is sent when the currently sent random action instruction is determined to be executed by the living body. 
 
     
     
       14. The non-transitory computer readable storage medium according to  claim 13 , wherein the random action instruction sequence comprises at least two categories of action instructions selected from a group comprising a first category of action instructions, a second category of action instructions and a third category of action instructions; and
 wherein the program instructions are being executed to cause the computer or the processor to perform at least one step of:
 for each first category of action instructions, determining whether a currently sent action instruction of the first category is executed by the living body according to an image captured by the image sensor; 
 for each second category of action instructions, determining whether a currently sent action instruction of the second category of is executed by the living body according to information detected by the at least one non-image sensor; and 
 for each third category of action instructions, determining whether a currently sent action instruction of the third category is executed by the living body according to the image captured by the image sensor and the information detected by the non-image sensor. 
 
 
     
     
       15. The non-transitory computer readable storage medium according to  claim 13 ,
 wherein the first detection information comprises at least one of facial gesture, facial key points, image texture information, image luminance, and an object recognized in the image captured by the image sensor, 
 wherein the at least one non-image sensor comprises at least one of a light sensor, a distance sensor, an acceleration sensor and a gyroscope, and 
 wherein the second detection information comprises at least one of light intensity, distance information, acceleration information, and gyroscope information.